Policy Brief - EVALUATING GENDER BIAS IN AI APPLICATIONS USING HOUSEHOLD SURVEY DATA
Coherent Identifier 20.500.12592/bqxg2v

Policy Brief - EVALUATING GENDER BIAS IN AI APPLICATIONS USING HOUSEHOLD SURVEY DATA

21 August 2023

Summary

The Equitable AI Challenge - implemented through DAI's from the Ghana demographic Digital Froners - was designed to fund approaches that will increase the and health survey, along with satellite imagery, were used to accountability and transparency of AI systems used in global construct the module. [...] Gender-specific AI models are trained using the Wealth Index as the dependent variable, and a range of geospaal data from satellite Geo Data MMalael eD aDtata AI Models DHS Data Gender Classification Comparison Female Data AI Models Geo Data “ We find that imagery and other sources as the independent variable. [...] A valuable feature of the type of AI models used is the ability to measure the importance of specific features (the geospaal variables) used to esmate wealth in the models. [...] Understanding what drives the differences between models trained on male and female data is also important - can other geospaal data features used in model training improve the performance of female models? Future research might also explore the possibility of ulizing wealth indices other than those produced by the DHS. [...] APPROPRIATE? CAN FUTURE SURVEYS BE IMPROVED TO ASSESS GENDER- SPECIFIC WEALTH? 3 The views expressed are solely the responsibility of the authors and do not necessarily reflect the views of the development partners.

Pages
4
Published in
Ghana